Induction of intestinal pro-inflammatory immune responses by lipoteichoic acid

The intestinal immune system must co-exist with resident commensal microorganisms while maintaining the ability to defend against potential microbial challenge. This immune tolerance is a highly regulated process comprised of a myriad of biological checkpoints necessary to maintain homeostasis between the host and the gut microbiota [1]. In instances of inflammatory bowel disease (IBD), this tolerance between immune cells and intestinal bacteria is disrupted; however, causes of this tolerance breakdown have not yet been determined [2, 3]. Although the etiology of IBD is still unknown, exaggerated inflammation induced by activated innate immune cells via their interaction with the microbiota and their gene products, as well as infiltrating CD4⁺ IFNγ⁺ T cells, likely play key roles in uncontrolled inflammation and tissue destruction [4‐6]. Foxp3⁺ regulatory T cells (Tregs) also critically control intestinal inflammation [7] and significantly prevent colitis [8], suggesting a pivotal role for Tregs in intestinal immune homeostasis [9].

A fundamental challenge in preventing an imbalanced immune response is the understanding of how the host immune system distinguishes a pathogen from normal intestinal flora. One of the commensal microorganisms of the gut is L. acidophilus, which expresses unique s urface l ayer p roteins (Slps), including A, B, X, and abundant lipoteichoic acid (LTA). LTA is a zwitterionic glycolipid found in the cell wall of several Gram-positive bacterial strains, including L. acidophilus, which facilitates the adhesion, colonization, and invasion of cells by the bacteria [10, 11]. The best studied form of LTA is composed of a polyglycerophosphate chain that is tethered to the membrane via a glycolipid anchor [12]. Studies indicate that LTA shares many of the inflammatory properties of lipopolysaccharide (LPS) via interactions with Toll-like receptors (TLRs) [13‐16] which evoke diverse responses in innate cells through distinct signaling cascades [17].

Previously, we have demonstrated that deletion of the gene responsible for LTA biosynthesis in L. acidophilus NCFM diminishes this bacterium's capacity to stimulate the immune system; thereby suppressing pathogenic CD4⁺T cells in induced colitis [18, 19]. To further investigate the role of LTA in inflammation, we engineered the NCK2031 strain in order to evaluate the effects, if any, of altered s urface l ayer p rotein (Slp) expression on LTA-induced pro-inflammatory signals and colitis.

Generation of NCK2031 To generate an L. acidophilus NCFM isogenic mutant defective in all three slp genes (slpA, slpB, and slpX), the slpB (LBA0175) and slpX (LBA0512) genes were sequentially deleted in an NCFMΔupp background host (NCK1909) using the upp-based counterselective gene replacement system [20]. Subsequently, attempts to insertionally inactivate the slpA gene (LBA0169) were made within the ΔslpBX double mutant (NCK2030) using a pORI-based gene knockout system [21]. The resulting ΔslpBX strain was deficient in SlpB and SlpX, but due to genetic instability of the insertion vector, NCK2031 continued to express SlpA (data not shown). Subsequently, wild-type L. acidophilus (NCK56), NCK2031, or LTA-deficient NCK2025 were propagated in de Man, Rogosa, and Sharpe broth (MRS, Difco) at 37°C for 15 hrs. The concentration of each L. acidophilus strain was adjusted to 1 × 10⁹ CFU/ml based on OD600 readings that had previously been correlated with CFU numbers [22]. Each mouse was orally treated with 5 × 10⁸ CFU bacterial strain; therefore 500 μl of 1 × 10⁹ CFU/ml suspension was centrifuged, pelleted and then resuspended in 100 μl of PBS. Each mouse was then treated with 100 μl PBS (control group) or 5 × 10⁸ CFU (of either L. acidophilus strain) orally in100 μl sterile PBS. Fecal pellets were collected from before, during, and for up to 8 days after L. acidophilus oral treatments. Each fecal pellet was resuspended in PBS (1:10 dilution, w/v). The suspension was then serially diluted and plated onto MRS agar containing Em (2 μg/mL). The homogenized material was serially diluted and plated onto MRS agar containing Em (2 μg/mL). The daily average CFU of the L. acidophilus strains in mouse feces were determined. For in vitro stimulation, bone marrow or human monocyte derived DCs were stimulated at a 1:1, 1:10, 1:100 ratio with live NCK2031, NCK56, NCK2025, or S. aureus-LTA. Subsequently the ratio of 1:1 was chosen, as this did not overwhelmingly activate DCs and therefore did not result in cell anergy or apoptosis. For the oral gavage of mice, each mouse received 5 × 10⁸ CFU of NCK2031, NCK56, or NCK2025 in 100 μL of PBS.

Cell culture Immature human [22, 23] or murine DCs [24] were treated with live L. acidophilus strains cells at 1:1 or with S. aureus-LTA (50 μg/mL) for 24 hrs. L. acidophilus- or S. aureus-LTA treated and untreated DCs (5 × 10⁵) were then stained and analyzed by BD FACSCaliber or a multicolor FACSCanto. Cell supernatants were also collected and analyzed using cytometric bead array kits (BD Biosciences). At least 1 × 10⁴ gated events per condition were acquired. Analysis software (BD CellQuest) allowed for calculation of cytokine values in supernatants at pg/mL.

Immunofluorescence The colons of mice (n = 5/group) treated with live NCK2031, NCK56, NCK2025, S. aureus-LTA, or PBS alone were fillet-opened, rolled, and snap frozen at -80°C in Tissue-Tek O.C.T. (Sakura Finetek USA, Inc., Torrance, CA). Sections (5 μm) were cut, fixed in ice-cold methanol (-20°C) for 15 min and blocked with 1% BSA. Sections were then incubated overnight at 4°C with purified hamster anti-mouse CD11c (BD Biosciences) and rat anti-mouse IL-10 (BioLegend), and antibodies for IL-12, IFNγ, TNFα, CD8, and Foxp3, followed by washing twice with PBS and incubation with anti-hamster AlexaFluor 594 and anti-rat AlexaFluor 488 (Invitrogen) for 1 hr. Sections were then washed twice with PBS and incubated for 10 minutes with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI, Invitrogen), washed with PBS two times and mounted with anti-fade mounting medium, as described previously [25]. Images were acquired using TissueGnostics Tissue/Cell High Throughput Imaging and Analysis System and analyzed using ImageJ software.

Flow cytometry Groups of mice (5 mice/group) were inoculated orally with NCK2031, NCK56, NCK2025, or S. aureus-LTA (5 × 10⁸ CFU/100 μL of sterile PBS/mouse) or PBS alone. Mice were then sacrificed after 1, 3, or 7 days; single cells were isolated from the colons [18], MLNs, and spleens and stained with CD11c, CD11b, CD103, MHC II, CD40, CD80, CD86, F4/80, CD4, CD3, CD80, or CD44 [26]. Stained cells were then fixed, permeabilized, and stained with IL-12, TNFα, IFNγ or isotype antibodies. At least 1 × 10⁵ gated events per condition were acquired. Data were acquired with BD FACSCanto II and analyzed using Tree Star FlowJo software.

DSS-Induced Murine Colitis Groups of C57BL/6 mice (n = 5/group) were inoculated orally with NCK2031, NCK56, NCK2025, or (5 × 10⁸ CFU/100 μL PBS/mouse) for four consecutive days. In addition, a group of mice was treated with purified S. aureus- LTA (12.5 μg/μL) [27] for 5 days. All of these mice received one 5-day cycle of 3% DSS in drinking water, followed by 3 days of regular drinking water, with or without S. aureus- LTA, and then sacrificed on day 13. Acute colitis was observed after the first cycle of DSS in the non-inoculated group. Disease progression, including weight loss, diarrhea and fecal hemoccult blood positivity (FOB), was monitored throughout the study. Thereafter, mice were sacrificed and excised colon cross-sectional Swiss rolls were fixed in 10% formaldehyde and embedded in paraffin. Tissue sections (4 μm) were stained with hematoyxylin and eosin (H&E), and blindly scored, as described previously [28, 29]. The grading, based on a scale from 0 to 28, takes into account the degree of inflammatory infiltrate, the presence of erosion, ulceration, or necrosis, and the depth and surface extension of the lesion.

Excessive stimulation of intestinal innate cells, including DCs, with commensal bacteria and/or their gene products results in immune dysfunction that can generate uncontrolled inflammation leading to tissue destruction and colitis [39]. Precise cellular and molecular mechanisms of these induced inflammatory immune responses in IBD remain poorly understood. Data so far indicate that chronic intestinal inflammation coincides with elevated levels of pro-inflammatory cytokines (i.e., IL-12, TNFα) [40], and the differentiation/activation of pro-inflammatory DC-subsets and pathogenic CD4⁺T cells [41]. Accordingly, promising studies show that inhibition of detrimental signals induced by stimulatory bacterial products mitigates IBD progression [42‐49].

Together, our data show that LTA plays a critical role in the induction of the inflammatory responses and cannot prevent mice from developing colitis. Contrary to our findings, other studies have shown that lactobacilli-LTA induces regulatory signals (i.e., IL-10) via Erk1/2 signaling, resulting in anti-inflammatory mechanisms [30]. Other reports also indicate that commensal LTA protects mice from DSS-induced colitis and ameliorates impaired epithelial tight junctions [27, 50]. This discrepancy may lay in the methodological approaches and reagents that were used in these studies. Here we demonstrate that LTA and LTA expressing bacteria posses the ability to stimulate innate immune component cells, including DCs and macrophages, which in turn, can trigger pathogenic IFNγ-secreting CD4⁺ T cells in disease progression.

The role of IFNγ expressed in the colon is unclear, as it has been shown that this cytokine exerts protective features in different models of inflammation [51]. For example, in a murine model of multiple sclerosis, neutralization of IFNγ resulted in inflammatory immune responses [52]. Similarly, in a mouse model of IBD, IFNγ has been shown to suppress IL-23 [53], and collagen-induced arthritis is induced in the absence of IFNγ signaling [54]. However, in the models we have previously used for experimental DSS inducing epithelial injury (i.e., leaky mucosa) resulting in colitis or pathogenic CD4⁺CD45RB^high T cell transfer; IFNγ seems to be involved in the pathogenesis of the induced colitis. Additionally, IFNγ contributes to intestinal pathology and lethality in lipopolysaccharide-sensitization models of toxic shock syndrome (TSS) [55, 56]. Further studies are needed to clearly elucidate the role of this cytokine in colitis and the timing of its production during disease progression.

Our results show that in vitro and in vivo, NCK2031 expressing LTA and purified S. aureus-LTA induce intestinal immune activation resulting in the production of TNFα and IL-12 and the development of pro-inflammatory cells such as CD11c⁺TNFα⁺IL-12⁺ DCs and F4/80⁺ macrophages that may contribute to an undesired CD4⁺IFNγ⁺TNFα⁺ T cell pro-inflammatory response seen in colitis. Conversely, treatment with LTA-deficient NCK2025 elicited regulatory signals and induced less inflammation, as demonstrated here and previously [18, 19]. We have previously demonstrated the involvement of SlpA of L. acidophilus NCFM in the immune regulation of DC functions [57]. Studies are underway to better delineate the physiological role of SlpA from this bacterium in vivo. Finally, an acute colitis characterized by bloody diarrhea, ulcerations and infiltrations with granulocytes can be induced in mice that are exposed for several days to significantly dissolved DSS polymers in drinking water [58, 59]. As seen in Figure 6, we also could confirm such clinical signs when DSS was used to induce colitis. Additionally, our data clearly show not only the proinflammatory nature of LTA but also the significantly exacerbated disease progression in the colon of mice that received LTA in the drinking water. Thus, these data are in sharp contrast to the notion that LTA plays a regulatory role and possesses the ability to mitigate DSS-induced colitis, as described previously [27]. In summary, oral administration of DSS to mice induces an acute colitis, followed by a slow colonic regeneration of the epithelium with a concomitant exaggerated inflammation; furthermore, LTA did not reduce the levels DSS-induced colitis or promote the rapid healing of ulcerated and destroyed epithelial tissues in our studies.

Our data strongly suggest that directed alteration of two of the L. acidophilus-Slps did not ameliorate LTA-induced pro-inflammatory signals and subsequent colitis. Additionally, further studies are required to elucidate the role of L. acidophilus-Slps in the steady state of normal gut homeostasis and in diseases such as colitis and colon cancer where deregulated inflammation plays a critical role in disease progression. Finally, LTA did not mitigate DSS-induced colitis but induced further colonic inflammation.